Most marketplace pitches lead with demand: users, downloads, GMV. Those numbers are real, and they are also the easiest to buy. I have run strategy and growth at ZenCabs, an electric cab service in Jammu, and growth at Equip9, a construction equipment marketplace. These are the questions I care about, roughly in the order I would ask them.
Start with the side that is harder to get
In most marketplaces one side is harder to recruit and keep, and for ride-hailing it is drivers. A rider who opens the app to no cars rarely opens it again, so money spent on demand before supply is ready is money wasted. At ZenCabs we onboarded 43 driver partners, with recruitment, training and incentives designed from scratch, before pushing for riders.
- Which side is constrained today, and how do you know?
- What does it cost to add one unit of supply, and how long until it earns?
- What share of the supply active six months ago is still active?
- What happens to supply earnings if demand dips for a month?
Utilisation, and what each unit of supply earns
If the company owns or finances its supply, utilisation is the business. At ZenCabs, moving fleet utilisation from 40% to 75% took revenue per car to ₹83K a month, which no acquisition campaign could have matched with the same effort. I would ask for utilisation month by month, revenue per unit of supply, and whether adding supply lowers utilisation for everyone already on the platform.
Completion rate, the metric hiding in operations
Cancellations and failed transactions look like an operations issue and behave like a retention issue. Every cancelled ride costs more than the acquisition that produced it, because the customer you paid for is now less likely to return. ZenCabs cut driver-side cancellations from 40% to 6% with 10+ operating procedures across onboarding and dispatch, and a marketplace should report the two sides separately: rider cancellations there moved for a different reason, falling from 200 to 80 a day when we stopped sending cars to far-off zones.
The useful questions are what share of requests end in a completed transaction, which side cancels more and why, and how retention differs between customers whose first order went well and those whose first order failed.
Repeat usage by cohort
Total users is the least informative number in the deck. What matters is how many come back. At ZenCabs, 27% of the user base became 8,100 monthly active riders, and a founder's version of that number is the first thing I would ask for on demand. Monthly cohorts show more than a blended rate: whether usage flattens, declines or grows as each group ages.
Unit economics that survive the discounts ending
| Metric | Why it matters | What to check |
|---|---|---|
| GMV | Total value flowing through the platform | Whether growth comes from volume, price or subsidy |
| Take rate | The share of GMV the company keeps | Whether it is stable, and what stops it being competed down |
| Average order value | Revenue per transaction | Whether it is rising for real reasons or because of mix |
| Contribution per transaction | What is left after direct costs and incentives | Whether it is positive without discounts |
| Payback on acquisition | How fast a new customer repays their cost | Measured on retained customers, excluding one-time sign-ups |
Take rates deserve extra scrutiny in categories where competitors are changing the model. Indian ride-hailing moved toward flat driver subscriptions in 2024 and 2025, which turns a percentage commission into a software fee. I cover that shift in the unit economics of an electric cab fleet.
The numbers most often dressed up
- Downloads and sign-ups presented as users.
- GMV that includes heavy discounts which will stop when the subsidy does.
- Supply counted when onboarded, including partners who have since stopped working.
- Blended retention that lets a strong early cohort hide a weak recent one.
- Annualised run-rate taken from one strong month.
Run-rate is a fair measure of momentum when the month behind it is typical, so ask to see the months either side. ZenCabs' ₹4 Cr is a run-rate too, which is why on this site I show the monthly GMV of ₹33L+ and the 10,000+ rides behind it.
Who owns the supply, and who paid for it
For any marketplace that owns or finances its supply, the balance sheet belongs in the diligence alongside the product. BluSmart ran an owned electric fleet in three cities and stopped taking bookings in April 2025 after SEBI alleged that ₹262 crore of loans raised for 6,400 vehicles had been diverted. Most of its cars were owned by a separate listed company run by the same founders. The questions to ask are who legally owns the assets, who lent against them, and on what terms.
How the team runs the business
Finally, ask how the founders know what is happening. A team that reviews its main numbers every week, with owners and thresholds, finds problems in days. At ZenCabs a weekly KPI review with the founder was the most useful thing we built, and I describe it in the weekly operating review.
One more question tells me a lot: which slow, permanent change in the market worries you most? Founders who have thought about it answer at once. The phrase comes from the classify step of my Push and Absorb framework, where slow, permanent forces are the ones that do the most damage.
